Datenvisualisierung 3

Darstellung der zusammenfassenden Statistik

Daniela Palleschi

Humboldt-Universität zu Berlin

2024-01-24

Readings

Kapital 27 (A field guide to base R) in Wickham et al. (2023)
course website: Ch. 12: base R

Learning objectives

Today we will…

learn what base R is
compare base R and tidyverse
learn base R equivalents of tidyverse verbs

base R

basic software containing the R programming languag
- contains the base package which is required to run R
includes several packages such as utils and stats (among others)
- installed when you install R

tidyverse

the Tidyverse (Wickham et al., 2019a) is a family of R-packages designed to facilitate cleaning and wrangling data
- tidyverse packages “haben eine gemeinsame Designphilosophie und eine gemeinsame Grammatik und Datenstruktur, so dass das Erlernen eines Pakets das Erlernen des nächsten erleichtert.” (Wickham et al., 2019b). - tidyverse was written in the R programming language

base R vs. tidyverse

main goal of base R is stability
- not many or often changes to functionality of functions
the tidyverse is constantly adding, updating, and changing functions
this means that tidyverse code is prone to “breaking”: tidyverse code that runs today might not run in a few years if some functions or arguments have been “deprecated”

Controversy

some people strongly prefer to use base R or the tidyverse
- arguments for tidyverse: more human-readable, tidier, simpler for non-programmers
- arguments for base R: “truer” R-programming, more stable
generally, having a strong background in one and at least a basic literacy of the other is wise

Twitter debates

What a thing to say when modern R is pretty much synonymous with the tidyverse for many in the community!

I was a base R masochist once too.. but there's no need for statements like this when the tidyverse has helped so many of us be more productive and write more readable code.
— Bodo Winter ((BodoWinter?)) January 10, 2023

in this Tweet we see the original post from Prof. Zorn stating that knowing the tidyverse does not equate knowing R
- but there were many replies highlighting the benefits of the tidyverse
- from instructors, professors (like Bodo Winter, who wrote a statistics books for linguists using R (Winter, 2019)), and data scientists working in industry

My preference

I obviously think learning the tidyverse is important
- the tidyverse is human centred, and we are not programmers or computer scientists
not everybody agrees with me, but there are also a lot of people who do

Set-up

pacman::p_load(
  tidyverse,
  here
)

Read in data

we’ll now see our first comparison of tidyverse code versus base R code

CSV: tidyverse

tidyverse: readr::read_csv()

df_tidy <-
  read_csv(
    here("daten", "languageR_english.csv")
  )

CSV: base R

base R: read.csv()

df_base <-
  read.csv(
    here("daten", "languageR_english.csv")
  )

Comparing outcome

how many columns?

length(df_tidy)

[1] 7

length(df_base)

[1] 7

what are the column names?

names(df_base)

[1] "AgeSubject"       "Word"             "LengthInLetters"  "WrittenFrequency"
[5] "WordCategory"     "RTlexdec"         "RTnaming"

names(df_tidy)

[1] "AgeSubject"       "Word"             "LengthInLetters"  "WrittenFrequency"
[5] "WordCategory"     "RTlexdec"         "RTnaming"

how many rows?

nrow(df_tidy)

[1] 4568

nrow(df_base)

[1] 4568

the data structure is identical

Wrangling columns and rows

we’ll see base R alternatives to our most common dplyr verbs

Extract variables: tidyverse

tidyverse: dplyr::select()

df_tidy |> 
  select(AgeSubject)

# A tibble: 10 × 1
   AgeSubject
   <chr>     
 1 young     
 2 young     
 3 young     
 4 young     
 5 young     
 6 young     
 7 young     
 8 young     
 9 young     
10 young

Extract variables: base R

the dollar sign ($) can be used to extract a column from a dataframe (or tibble)
this will give us a vector, whereas dplyr::select() preserves the dataframe/tibble attributes of the column

base R: dataframe$column_name

df_base$AgeSubject

 [1] "young" "young" "young" "young" "young" "young" "young" "young" "young"
[10] "young" "young" "young" "young" "young" "young" "young" "young" "young"

Extract variables: base R

or we can use dataframe[row,column]
we can use the name of a column in quotation marks

base R: dataframe[,'column_name']

# using variable name
df_base[,"AgeSubject"]

 [1] "young" "young" "young" "young" "young" "young" "young" "young" "young"
[10] "young" "young" "young" "young" "young" "young" "young" "young" "young"

or we can give the index of the column, where 1 means the first column, 2 means the second column, and so on

base R: dataframe[,index]

# using variable index
df_base[,1]

 [1] "young" "young" "young" "young" "young" "young" "young" "young" "young"
[10] "young" "young" "young" "young" "young" "young" "young" "young" "young"

Multiple variables: tidyverse

tidyverse: dplyr::select()

df_tidy |> 
  select(AgeSubject, RTlexdec)

# A tibble: 10 × 2
   AgeSubject RTlexdec
   <chr>         <dbl>
 1 young          695.
 2 young          600.
 3 young          547.
 4 young          617.
 5 young          633.
 6 young          687.
 7 young          584.
 8 young          527.
 9 young          741.
10 young          536.

Multiple variables: baseR

for this we need c()

base R: dataframe[,c('column_name1','column_name2')]

# using variable name
df_base[,c("AgeSubject", "RTlexdec")]

   AgeSubject RTlexdec
1       young   694.89
2       young   600.40
3       young   547.27
4       young   616.60
5       young   633.08
6       young   686.75
7       young   584.40
8       young   526.82
9       young   741.48
10      young   536.38

base R: dataframe[,c(index1,index2)]

# using variable index
df_base[,c(1, 6)]

   AgeSubject RTlexdec
1       young   694.89
2       young   600.40
3       young   547.27
4       young   616.60
5       young   633.08
6       young   686.75
7       young   584.40
8       young   526.82
9       young   741.48
10      young   536.38

Extract/Filter observations: tidyverse

with the filter() function from dplyr

tidyverse: dplyr::filter()

df_tidy |> 
  filter(RTlexdec > 600 & RTnaming < 480)

# A tibble: 856 × 7
   AgeSubject Word  LengthInLetters WrittenFrequency WordCategory RTlexdec
   <chr>      <chr>           <dbl>            <dbl> <chr>           <dbl>
 1 young      doe                 3             3.91 N                695.
 2 young      pork                4             5.02 N                617.
 3 young      prop                4             4.77 N                687.
 4 young      arc                 3             4.89 N                741.
 5 young      tile                4             4.08 N                647.
 6 young      slope               5             5.80 N                633.
 7 young      pith                4             2.48 N                696.
 8 young      blitz               5             4.19 N                672.
 9 young      port                4             6.08 N                683.
10 young      plan                4             7.46 N                636.
# ℹ 846 more rows
# ℹ 1 more variable: RTnaming <dbl>

Extract/Filter observations: base R

add these conditional statements into [,]
- we need to include the dataframe name with the dollar sign preceding the column name

base R: dataframe[,]

df_base[df_base$RTlexdec > 600 & df_base$RTnaming < 480,]

   AgeSubject  Word LengthInLetters WrittenFrequency WordCategory RTlexdec
1       young   doe               3         3.912023            N   694.89
4       young  pork               4         5.017280            N   616.60
6       young  prop               4         4.770685            N   686.75
9       young   arc               3         4.890349            N   741.48
17      young  tile               4         4.077537            N   647.07
18      young slope               5         5.802118            N   632.54
22      young  pith               4         2.484907            N   695.86
26      young blitz               5         4.189655            N   671.59
29      young  port               4         6.084499            N   683.36
34      young  plan               4         7.462789            N   636.10
   RTnaming
1     466.4
4     460.3
6     477.1
9     453.8
17    459.3
18    476.2
22    473.3
26    469.5
29    459.3
34    470.4

Select single data points: tidyverse

use filter() and select() (which we’ve already done before)

tidyverse: filter() |> select()

df_tidy |> 
  filter(RTlexdec > 600, RTnaming < 480) |> 
  select(AgeSubject, RTlexdec)

# A tibble: 10 × 2
   AgeSubject RTlexdec
   <chr>         <dbl>
 1 young          695.
 2 young          617.
 3 young          687.
 4 young          741.
 5 young          647.
 6 young          633.
 7 young          696.
 8 young          672.
 9 young          683.
10 young          636.

Select single data points: base R

combine row and column values in [,]

base R: dataframe[conditions,columns]

df_base[df_base$RTlexdec > 600 & df_base$RTnaming < 480,c("AgeSubject", "RTlexdec")]

   AgeSubject RTlexdec
1       young   694.89
4       young   616.60
6       young   686.75
9       young   741.48
17      young   647.07
18      young   632.54
22      young   695.86
26      young   671.59
29      young   683.36
34      young   636.10

Select single data points: base R

again, you can replace the column names with the index value

base R: dataframe[conditions,index]

df_base[df_base$RTlexdec > 600 & df_base$RTnaming < 480,c(1, 6)]

   AgeSubject RTlexdec
1       young   694.89
4       young   616.60
6       young   686.75
9       young   741.48
17      young   647.07
18      young   632.54
22      young   695.86
26      young   671.59
29      young   683.36
34      young   636.10

Create new variables: tidyverse

with the mutate() function from dplyr

tidyverse: dplyr::mutate()

df_tidy |> 
  mutate(rt_lexdec_s = RTlexdec/1000)

# A tibble: 4,568 × 8
   AgeSubject Word   LengthInLetters WrittenFrequency WordCategory RTlexdec
   <chr>      <chr>            <dbl>            <dbl> <chr>           <dbl>
 1 young      doe                  3             3.91 N                695.
 2 young      whore                5             4.52 N                600.
 3 young      stress               6             6.51 N                547.
 4 young      pork                 4             5.02 N                617.
 5 young      plug                 4             4.89 N                633.
 6 young      prop                 4             4.77 N                687.
 7 young      dawn                 4             6.38 N                584.
 8 young      dog                  3             7.16 N                527.
 9 young      arc                  3             4.89 N                741.
10 young      skirt                5             5.93 N                536.
# ℹ 4,558 more rows
# ℹ 2 more variables: RTnaming <dbl>, rt_lexdec_s <dbl>

Create new variables: tidyverse

define the new variable name (with dataframe$variable) and assign the value with the assignment operator <-

base R

df_base$rt_lexdec_s <- df_base$RTlexdec/1000

Summarise: tidyverse

summarise() from dplyr

tidyverse: dplyr::summarise()

  df_tidy |> 
  summarise(
    mean_lexdec = mean(RTlexdec),
    sd_lexdec = sd(RTlexdec),
    mean_naming = mean(RTnaming, na.rm = T),
    sd_naming = sd(RTnaming, na.rm = T)
  )

# A tibble: 1 × 4
  mean_lexdec sd_lexdec mean_naming sd_naming
        <dbl>     <dbl>       <dbl>     <dbl>
1        708.      115.        566.      101.

Summarise: tidyverse

we have to create new objects containing the value of each operation + combine them into a data frame using the data.frame() function
there are many alternative ways to do this, but this is the simplest if we only want to produce a few summary statistics

base R

data.frame(mean_lexdec = mean(df_base$RTlexdec),
           sd_lexdec = sd(df_base$RTlexdec),
           mean_naming = mean(df_base$RTnaming, na.rm = T),
           sd_naming = sd(df_base$RTnaming, na.rm = T))

  mean_lexdec sd_lexdec mean_naming sd_naming
1    708.1336  114.8599    565.9233  100.8153

Plots

ggplot2 is popular even among people who don’t use the tidyverse + this is because it has some useful features and a clean look

Plots: base R

can be useful when you want to produce simple plots to get a first look at your data
- the most useful functions are hist() and plot()
- note that these functions work with vectors, which is why we have to use $ to extract the columns from the data frame

hist(df_base$RTlexdec)

plot(df_base$RTlexdec, df_base$RTnaming)

Plots: tidyverse

as we’ve seen before:

library(patchwork)

# histogram
fig_hist <- 
  df_base |> 
  ggplot() +
  aes(x = RTlexdec) +
  geom_histogram()

# scatter plot
fig_scatter <-
  df_base |> 
  ggplot() +
  aes(x = RTlexdec, y = RTnaming) +
  geom_point() 
  
fig_hist + fig_scatter

Abbildung 1: Histogram and scatterplot with ggplot2

Learning objectoves 🏁

In this chapter we will…

learn what Base R is ✅
compare Base R and Tidyverse ✅
get to know the Base R equivalents of Tidyverse verbs ✅

Tasks

In the Console

rm(list=ls())
.rs.restartR()

Convert the following tidyverse code to base R. We will again use the languageR_english.csv dataset.

Read-in

df_eng <- 
  read_csv(here("daten", "languageR_english.csv"))

Selecting columns

df_eng |> 
  select(Word, WrittenFrequency)

# A tibble: 10 × 2
   Word   WrittenFrequency
   <chr>             <dbl>
 1 doe                3.91
 2 whore              4.52
 3 stress             6.51
 4 pork               5.02
 5 plug               4.89
 6 prop               4.77
 7 dawn               6.38
 8 dog                7.16
 9 arc                4.89
10 skirt              5.93

Filtering rows

df_eng |> 
  filter(WrittenFrequency > 5.6)

# A tibble: 10 × 7
   AgeSubject Word   LengthInLetters WrittenFrequency WordCategory RTlexdec
   <chr>      <chr>            <dbl>            <dbl> <chr>           <dbl>
 1 young      stress               6             6.51 N                547.
 2 young      dawn                 4             6.38 N                584.
 3 young      dog                  3             7.16 N                527.
 4 young      skirt                5             5.93 N                536.
 5 young      are                  3            11.3  N                611.
 6 young      pipe                 4             6.00 N                563.
 7 young      guard                5             6.59 N                559.
 8 young      slope                5             5.80 N                633.
 9 young      pile                 4             6.16 N                595.
10 young      tide                 4             6.08 N                598.
# ℹ 1 more variable: RTnaming <dbl>

Selecting rows and columns

df_eng |> 
  filter(WrittenFrequency > 5.6 & AgeSubject == "old") |> 
  select(AgeSubject, Word, WrittenFrequency)

# A tibble: 10 × 3
   AgeSubject Word   WrittenFrequency
   <chr>      <chr>             <dbl>
 1 old        stress             6.51
 2 old        dawn               6.38
 3 old        dog                7.16
 4 old        skirt              5.93
 5 old        are               11.3 
 6 old        pipe               6.00
 7 old        guard              6.59
 8 old        slope              5.80
 9 old        pile               6.16
10 old        tide               6.08

Scatterplot

df_eng |> 
  ggplot() +
  aes(x = WrittenFrequency, y = RTlexdec) +
  geom_point()

Tidyverse versus base R

What is your impression of base R versus the tidyverse? Based on what you’ve seen, would you prefer one over the other, or would you prefer one in certain cases only? There’s no correct answer here.

Session Info

Hergestellt mit R version 4.3.0 (2023-04-21) (Already Tomorrow) und RStudioversion 2023.9.0.463 (Desert Sunflower).

print(sessionInfo(),locale = F)

R version 4.3.0 (2023-04-21)
Platform: aarch64-apple-darwin20 (64-bit)
Running under: macOS Ventura 13.2.1

Matrix products: default
BLAS:   /Library/Frameworks/R.framework/Versions/4.3-arm64/Resources/lib/libRblas.0.dylib 
LAPACK: /Library/Frameworks/R.framework/Versions/4.3-arm64/Resources/lib/libRlapack.dylib;  LAPACK version 3.11.0

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] patchwork_1.1.3 janitor_2.2.0   here_1.0.1      lubridate_1.9.2
 [5] forcats_1.0.0   stringr_1.5.0   dplyr_1.1.3     purrr_1.0.2    
 [9] readr_2.1.4     tidyr_1.3.0     tibble_3.2.1    ggplot2_3.4.3  
[13] tidyverse_2.0.0

loaded via a namespace (and not attached):
 [1] utf8_1.2.3       generics_0.1.3   stringi_1.7.12   hms_1.1.3       
 [5] digest_0.6.33    magrittr_2.0.3   evaluate_0.21    grid_4.3.0      
 [9] timechange_0.2.0 fastmap_1.1.1    rprojroot_2.0.3  jsonlite_1.8.7  
[13] fansi_1.0.4      scales_1.2.1     cli_3.6.1        rlang_1.1.1     
[17] crayon_1.5.2     bit64_4.0.5      munsell_0.5.0    withr_2.5.0     
[21] yaml_2.3.7       tools_4.3.0      parallel_4.3.0   tzdb_0.4.0      
[25] colorspace_2.1-0 pacman_0.5.1     vctrs_0.6.3      R6_2.5.1        
[29] lifecycle_1.0.3  snakecase_0.11.0 bit_4.0.5        vroom_1.6.3     
[33] pkgconfig_2.0.3  pillar_1.9.0     gtable_0.3.4     glue_1.6.2      
[37] xfun_0.39        tidyselect_1.2.0 rstudioapi_0.14  knitr_1.44      
[41] farver_2.1.1     htmltools_0.5.5  rmarkdown_2.22   labeling_0.4.3  
[45] compiler_4.3.0

Literaturverzeichnis

Wickham, H., Averick, M., Bryan, J., Chang, W., McGowan, L. D., François, R., Grolemund, G., Hayes, A., Henry, L., Hester, J., Kuhn, M., Pedersen, T. L., Miller, E., Bache, S. M., Müller, K., Ooms, J., Robinson, D., Seidel, D. P., Spinu, V., … Yutani, H. (2019a). Welcome to the tidyverse. Journal of Open Source Software, 4(43), 1686. https://doi.org/10.21105/joss.01686

Wickham, H., Averick, M., Bryan, J., Chang, W., McGowan, L., François, R., Grolemund, G., Hayes, A., Henry, L., Hester, J., Kuhn, M., Pedersen, T., Miller, E., Bache, S., Müller, K., Ooms, J., Robinson, D., Seidel, D., Spinu, V., … Yutani, H. (2019b). Welcome to the Tidyverse. Journal of Open Source Software, 4(43), 1686. https://doi.org/10.21105/joss.01686

Wickham, H., Çetinkaya-Rundel, M., & Grolemund, G. (2023). R for Data Science (2. Aufl.).

Winter, B. (2019). Statistics for Linguists: An Introduction Using R. In Statistics for Linguists: An Introduction Using R. Routledge. https://doi.org/10.4324/9781315165547